Tagan TG480-U01 480W ATX PSU

[JEWilson]

I rec'd this PSU dead

The O/P caps (Fucku) were vented, replaced all of these with
Pana FC and Samxon GD.

Post recap. The PSU is still dead.

Fuse ok
Bridge rectifier ok.
Removed the primary HS and tested the ICs - all ok.

However, I found one power resistor rated at 100R 5%, reading 46.
The pcb was slightly brown and this resistor will have to be replaced.
Also, another resistor 1.8R is reading more than 5% out of spec., that
is. 1.7R with my ESR meter. Ok, 1.8R 5% (+/- 0.09) is there or there
abouts. However, as this had been shrouded, the shroud had welded
itself to the device. Perhaps it is starting to drift - so will replace.

I'm working my way through this PSU to see what else is wrong with
it (if anything).

Just one question, these resistors are power resistors.
I believe these may 1 to 2W wirewound ceramic but not ceramic
rect. block power devices.

What should I be looking for, quality wise?

Any advice - appreciated

Thx.

[JEWilson]

Re: Tagan TG480-U01 480W ATX PSU

Replaced the resistor with one recovered from a dead PSU

The secondary power rectifiers are all good.

The PSU is still dead.
However, it make a silght noise, like a ticking sound as if
starting into a dead short.

Suspect the controller IC (TL3842P) or the PFC corrector IC (STL6561).

[i4004]

Re: Tagan TG480-U01 480W ATX PSU

how can resistor's resistance go down?
i think only failure mode is it going up...

check those chips by seeing if supply pin is shorted to ground(prior to that make sure it's not connected to rest of the circuit...)

if it's not them, unhook everything from secondary and attach some dummy load...

ticking indeed sounds like short circuit on secondary...

[PCBONEZ]

Re: Tagan TG480-U01 480W ATX PSU

how can resistor's resistance go down?

It's called melting.
Light bulbs can short too.
Not impossible, just rare.

.

[JEWilson]

Re: Tagan TG480-U01 480W ATX PSU

Y - seems odd - resistance going down but had, as stated, a slight
browning to the pcb. This supply is quite crowded - sorry no pics.

I should emphasise the sound is difficult to discern it's location exactly.
It is barely audible and attempting to locate it, watching my ears of course,
is a problem.

However, the supply has no +5VSB.
I will check the rectifier diode at the O/P for this rail
and replace the opto-isolator as a matter of course.
If this supply has, as is normally the case, a TL431C.
I may replace this also. But do these fail often?

The supply is getting AC as I can measure across the bridge rectifier and
get a reading.

Vcc is non-existent to the PFC IC, I cannot get access to the PWM controller
given this sub-pcb is in the way.

Fix the +5VSB first.

Thx

[i4004]

Re: Tagan TG480-U01 480W ATX PSU

yes, +5vstb first, offcourse.

>It's called melting.
Light bulbs can short too.
Not impossible, just rare

if resistor is melting it's turning into coal, or something simillar.
and is that a good conductor?
and if it's wirewound i also can't see a way for it to go down.

[PCBONEZ]

Re: Tagan TG480-U01 480W ATX PSU

There are several different kinds of cores used in resistors and several of them are based on metal oxides.

[JEWilson]

Re: Tagan TG480-U01 480W ATX PSU

Y - perhaps I should clarify

Upon close inspection, that resistor does not look w/wound.
My assumption is that it is thus a metal oxide power resistor.

Well.... regards +5VSB

I have replaced;
The opto-isolator - no change to PSU state;
Checked the rectifire diode - this is good.
Replacd the TL431 - no change to PSU state.
The lytics were changed at recap to 2 of 1000uF 10V Pana FC.

So... as to the PFC sub-pcb;
Next will remove the diodes, power resistors et al and chk. values.
I did replace two lytics there (on the sub pcb). These are for
power to the IC. Replaced with 33uF 50V.

[i4004]

Re: Tagan TG480-U01 480W ATX PSU

>and several of them are based on metal oxides.

offcourse..
so what happens to metal oxide when it burns?
should i just put one on fire and see what happens?

[JEWilson]

Re: Tagan TG480-U01 480W ATX PSU

There might exist an alternative answer to the matter of the
resistor. Was it incorrectly marked and otherwise ok?

I have, only on rare occassions seen this to be so.

[i4004]

Re: Tagan TG480-U01 480W ATX PSU

hm..was there some other residue present, glue, or whatever?
perhaps it made conductance rise.

i should really burn one metal film resistor!
what's better than experiment?

[JEWilson]

Re: Tagan TG480-U01 480W ATX PSU

None

[Brian C]

Re: Tagan TG480-U01 480W ATX PSU

I believe that 1.8R resistor is current sensing resistor for primary switch. It should be high precision power resistor. Please check the DC voltage across primary reservoir cap, should be somewhere around 350Vdc. The active PFC is OK if the voltage reading is correct.

[PCBONEZ]

Re: Tagan TG480-U01 480W ATX PSU

>and several of them are based on metal oxides.

offcourse..
so what happens to metal oxide when it burns?
should i just put one on fire and see what happens?

Metal ~OXIDE~ is already oxidized and therefore already 'burned'.
However if you heat it enough you will produce a pool of metal with slag floating on top.

.

[JEWilson]

Re: Tagan TG480-U01 480W ATX PSU

I shorted one of a pair of 2SK2837 MOSFETs on the PFC sub-pcb
accidentally. This blew the mains fuse.

I purchased some of these and replaced both of them on the
PFC sub pcb. Post replacement, I replaced also the ST L6561
PFC IC as well as a 27uF 400V cap.

This supply is an active PFC design.

The PFC sub pcb is normally mounted via screws to the primary and
secondary heatsinks in the PSU in an upside down fashion.

To locate the source of the ticking sound, I removed the
sub pcb from the heatsinks and sat in on insulated foam.

The tick sound clearly comes from the PFC sub pcb but it
is still difficult to locate it exactly.

The sub pcb is attached to the PSU by means of a 4 pin
header. This sources AC from the primary after mains
filtration and noise suppresion and returns, I believe, DC +/- to
the PSU primary (on the main PSU body) after AC rectification and
phase correction by the PFC IC.

Note, the AC bridge is located on the PFC sub pcb.
There is also one of 27uF 400V cap (replaced as above).
Both of the large Pana UP caps are in the main body
of the PSU primary side.

If as I believe, the noise is consistent with starting into
a short, being as I have replaced both of the 2SK2837s,
I can only consider the AC bridge as being at fault in
cct., as it tests ok out of cct.

Would this be a correct assessment of this topology?

[i4004]

Re: Tagan TG480-U01 480W ATX PSU

i don't understand your troubleshooting technique at all.
you seem to be doing a lot of desoldering, which i think you should avoid, and instead concentrate on measuring voltages and scoping, if the have the scope.

now, what is with that 5vstby?
is it done with a small pwm ic?
does it receive supply voltage?
brian said pfc is ok if voltage reading across it is correct, so what is it?
bridge rectifier?
well if it's not ok, it would skew the rectified voltage, right?
if it was shorted it would blew the mains fuse.

about pfc; grab datasheet for the chip you replaced, and check the electric diagrams inside: it probably has your "version" too, so check its voltages, esp. if voltage on main cap is wrong.

btw. i didn't quite get this from your 1st post
>Removed the primary HS and tested the ICs - all ok.
what do you mean?

what does primary heatsink(?) has to do with all other ics there?
sometimes i don't get your abbreviations at all.

oh yeah, i think some pix would come handy...

[Gianni]

Re: Tagan TG480-U01 480W ATX PSU

>Removed the primary HS and tested the ICs - all ok. what do you mean?

I think he has to remove HS to reach/check the ICs.
I agree with you (i4004), to discover what is wrong the best way is to check with a scope the voltages and waveforms according to ICs datasheet.
With a little patience, datasheet and following tracks, you can figure out what you should have on output or on components.

Also, another resistor 1.8R is reading more than 5% out of spec., that is. 1.7R with my ESR meter. Ok, 1.8R 5% (+/- 0.09) is there or there abouts.

Is your ESR meter a four wire meter?
My DMM is not so precise and if I measure 1.7R instead of 1.71R I would not think the resistor is going bad.

If this supply has, as is normally the case, a TL431C. I may replace this also. But do these fail often?

As far as I have seen they rarely fail.

Ciao
Gianni

[linuxguru]

Re: Tagan TG480-U01 480W ATX PSU

If you can bypass the A-PFC board, remove it. Connect the fused, EMI-filtered mains supply directly to the primary side rectifier bank, after verifying that the caps, rectifier and primary switchers are rated high enough to handle the rectified primary HV.

In my experience, the root cause of unreliability in many off-line PSUs is improperly-designed A-PFC. OTOH, PPFC is generally bulletproof.

[JEWilson]

Re: Tagan TG480-U01 480W ATX PSU

i4004

Note, I do not have a scope but would like to get one soon.

Thanks for your comments.

As the PSU was dead, I tested if any O/P shorted to gnd. Nope
I tested in cct. with MMeter, if ICs in the main PSU (not A-PFC) getting Vcc, they are.
Also checked if Vcc on these ICs shorted to gnd. No, they are not.

You may recall, I sought to establish why +5VSB not functioning.
As the PSU appeared completely dead, in the +5VSB cct., I tested
the following, first in cct. and then out of cct.
1. The opto-isolator, a KB817B appeared ok
2. The +5VSB rectifier diode, appeared ok
3. The TL431, seemed ok

In any event, I replaced both the TL431 and the KB817B with new parts.
Result : +5VSB still not functioning.

Next, tested the primary section and specifically, the start up resistors.
These all appear to be correct values (out of cct.) with the exception
of the resistor that was wrong (see prior post).
Replaced this resistor with a known good device at correct value.

Established the PSU was making a small ticking sound.
Difficult to establish where it was coming from.
Ticking sound consistent with starting into short.
i4004 advised this may be in the secondary, consistent with fault condition.

To better isolate location of sound, removed A-PFC from PSU unit as is typically mounted.
This is simply two screws connecting the A-PFC to the priamry and secondary heatsinks.
The A-PFC is connected to the main PSU by a 4 pin jumper header.
This supplies AC from the main PSU, 2 pins.
The A-PFC sends back phase corrected V+ and V- to the main PSU primary and smoothing.
Would this be correct?

The A-PFC sub pcb is getting mains AC voltage, MMeter reads 245V AC across bridge rectifier.
The brige rectifier is located on the A-PFC pcb.
Surmise, AC supply, post mains filtration supplied by main PSU (from outlet) is ok.

Note however, the sound does indeed issue from the A-PFC pcb.
Tested voltages across the ST L6561 PFC IC.
Device gets Vcc and is not shorted to gnd.
I have looked at the L6561 datasheet and further tested the other pins
on the device for readings with MMeter in cct., these read;
Pin 1, Err Amp I/P, reads : 2.49-2.51V
Pin 2, Err Amp O/P, reads : 2.22-2.41V
Pin 3, Mult - Voltage proportional to the rectified mains, appears on this pin, reads : 2.67-2.79V
Pin 4, Current Sense, reads : 0.01-0.02mV
Pin 5, Zero Current Detect, reads : 0.02V-1.09V
Pin 6, GND, reads : 0.00V
Pin 7, Gate Drive reads : 0.01-73.5mV
Pin 8, Vcc, reads : 12.48-16.08V

As detailed prior to, shorted one of the 2SK2387 MOSFETs, blew mains fuse at the outlet.
This no suprise as this is what is meant to happen.
In any event, replaced these with two new devices.

Suspected the L6561 IC was faulty (the ticking sound) as has been suggested by others.
Replaced this part with a known good L6561.
The A-PFC is still making a ticking sound.

As the L6561 IC has been replaced as well as the power ICs as with;
2 of Toshiba 2SK2837, given further; 1 of WTE U1006FCT 10A Fast Recovery Rectifier tests ok in/out of cct.
As a result, PFC IC known good.

Diagnoses;
One or more of the discretes around the PFC IC may be out of tolerance, open or s/c.
This being the case, no surprise the supply is not starting i.e. the A-PFC is in error and how, in turn, the PSU has no +5VSB.

To these ends,
Have tested all the diodes in this section. Out of cct., these all test ok with MMeter as with;

D2 - fwd (red to blk) A -> C, reads 0.515, fwd (blk to red) A -> C, reads infinite, ok
D3 - fwd (red to blk) A -> C, reads 0.747, fwd (blk to red) A -> C, reads infinite, ok
Has a 100kpF ceramic disc cap in shunt, no s/c
D4 - fwd (red to blk) A -> C, reads 0.604, fwd (blk to red) A -> C, reads infinite, ok
D5 - fwd (red to blk) A -> C, reads 0.604, fwd (blk to red) A -> C, reads infinite, ok
D6 - fwd (red to blk) A -> C, reads 0.508, fwd (blk to red) A -> C, reads infinite, ok

Have sketched out the A-PFC cct. for fault tracing
Have tested the power resistors on the A-PFC as with;

R10 - Blk Grn Gry Gld Grn = 0, 5, 8 x 0.1 = 0.058R 0.5% (+/- 0.00029)
Reads out of cct - 0.5R with ESR/Low Ohms Meter, appears ok
Gianni, note, using Bob Parker ESR meter.

R1 - Red Red Yel Gld = 2, 2 x 10k = 220.0kR 5% (+/- 11000)
Reads out of cct - 216kR with MMeter, appears ok

R14 - Org Whi Blk Gld = 3, 9 x 1 = 39.0 R 5% (+/- 1.95)
Reads out of cct - 38.6R with MMeter, appears ok

Linuxguru - your suggestion as to bypassing the A-PFC is sound.
Not yet sure if I can do this given the bridge rectifier is on the A-PFC pcb.

To Do
Test the voltage across primary smoothing caps as Brian C suggested.
If ok, test the current at pins 1, 2, 3, 4, 5 and 7 of L6561.
Test the remainder of the discretes in the A-PFC section.

Apologies if prior post(s) have been less than clear

[i4004]

Re: Tagan TG480-U01 480W ATX PSU

>Test the voltage across primary smoothing caps as Brian C suggested.

yes, that first.


>The A-PFC sends back phase corrected V+ and V- to the main PSU primary and smoothing.
Would this be correct?

yes. it also means you can follow linuxguru's suggestion about bypassing the apfc.
about bridge rectifier being on apfc board; use that stuff from some dismantled psu, or so; you only need fuse, emi filter, bridge rectifier and that big cap.
even those chinese crap psus you can use to isolate those portions and bring that 350vdc voltage to your target.
or just take bridge off the apfc board and connect it to prior portions of the circuit on one side, and primary cap on another side.
if you do that keep everything well isolated. mark what was where(prior to taking it off) so you don't make mistakes.

OFFCOURSE, first measure voltages so you're sure apfc is not working, though by now it seems almost certain it's malfunctioning.

>This being the case, no surprise the supply is not starting i.e. the A-PFC is in error and how, in turn, the PSU has no +5VSB.

to explain a bit of how 5vstby usually works today:
mains voltage is rectified and smoothed(you know how...with apfc it's more complex, but in the end you have 350Vdc) and then it first feeds the 5vstby pwm ic(which in turn feeds the +5vstby trafo, and then trafo makes the +5vstby voltage that supplies the MAIN pwm ic(one that steers the main trafo)): NOW, i think you never mentioned that chip, because tl431 is not it

that's the shunt, not pwm ic.
these chips are usually 8 pin dil that take +350vdc and supply the 5vstby trafo, ergo they don't need external big fets, as the power they provide(to trafo) is typically low.
it can also be 3 pin topswitch, tinyswitch or alike.

now, can you locate that chip, what is it, and is it getting +350v?
if not, that's one more piece of puzzle that points to apfc, and ergo one more reason to bypass it.

oh yeah, when working with mains(or rectified mains) and looking at it, i think you may wanna wear safety goggles, and don't come too near.
weird and explosive things can happen(even when you don't short the fet..
<wink> )

safety first, all other shit later.

oh yeah, again, would it be possible we see that apfc circuit, so we can perhaps speculate what's making it tick...heh